1; RUN: opt -mtriple=s390x-unknown-linux -mcpu=z13 -loop-vectorize \
2; RUN:   -force-vector-width=2 -debug-only=loop-vectorize \
3; RUN:   -disable-output < %s 2>&1 | FileCheck %s
4; REQUIRES: asserts
5;
6; Check that a scalarized load does not get operands scalarization costs added.
7
8define void @fun(i64* %data, i64 %n, i64 %s, double* %Src) {
9entry:
10  br label %for.body
11
12for.body:
13  %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
14  %mul = mul nsw i64 %iv, %s
15  %gep = getelementptr inbounds double, double* %Src, i64 %mul
16  %bct = bitcast double* %gep to i64*
17  %ld = load i64, i64* %bct
18  %iv.next = add nuw nsw i64 %iv, 1
19  %cmp110.us = icmp slt i64 %iv.next, %n
20  br i1 %cmp110.us, label %for.body, label %for.end
21
22for.end:
23  ret void
24
25; CHECK: LV: Found an estimated cost of 2 for VF 2 For instruction:   %mul = mul nsw i64 %iv, %s
26; CHECK: LV: Found an estimated cost of 2 for VF 2 For instruction:   %ld = load i64, i64* %bct
27}
28